Unintended pregnancy is an issue of substantial personal and social concern in the United States. Efforts to reduce its level center around improving contraceptive use among sexually active couples at risk of unintended pregnancy. The proposed project would increase our knowledge about the relationship between contraceptive method use and unintended pregnancy. We would estimate one important indicator of success among American couples using contraceptives--the proportions of method users who become accidently pregnant while using a method (use-failure rates). We would determine the extent to which this indicator differs across contraceptive methods and across subgroups of users, as well as investigate changes between the mid- 1980s and early 1990s and the trends in use-failure rates over 6-, 12- and 24-month periods of use. For the U.S. as a whole we would estimate total and sociodemographic subgroup- specific levels of pregnancy and unintended pregnancy and analyze components of change over time, including the contribution of changes in contraceptive use patterns and effectiveness. To accomplish these goals, we would field a survey of a nationally representative sample of approximately 9,500 women having abortions in order to gather information on contraceptive use at time of conception and on other sociodemographic characteristics unavailable from other sources. We would use these data, together with already available national abortion statistics, to assess the accuracy of abortion reporting in the 1995 National Survey of Family Growth (NSFG). If unintended pregnancies that ended in abortion are underreported in the 1995 NSFG, as they have been in all past NSFGs and other national surveys of women's reproductive behavior that have been examined, we would use procedures developed under a prior NICHD grant to correct NSFG data for abortion underreporting and to calculate more accurate measures of contraceptive use-failure than would be possible from the NSFG alone. Using these corrected estimates, we would examine sociodemographic characteristics of women associated with contraceptive success. Findings from this project would contribute to monitoring the Year 2000 Objectives and help direct their achievement; they would assist those trying to choose among contraceptive methods and those seeking to provide them with accurate information; and, they would provide assistance to other researchers in the use of NSFG data for analysis of pregnancy and pregnancy outcomes.